Reflecting on 30 years of miRNA biology in malignant hematology: current challenges and future directions Open Access

The discovery of microRNAs (miRNAs), a class of noncoding RNAs ∼21 to 25 nucleotides in length, in the early 1990s marked a significant turning point in our understanding of molecular biology and gene regulation. Pioneering work by Victor Ambros¹ and Gary Ruvkun,² who were awarded with the Nobel Prize in Physiology or Medicine in 2024, demonstrated the negative regulation of lin-14 expression through the interaction of lin-4 RNAs to conserved elements in the 3′ untranslated region (3′ UTR). This significant observation first highlighted the temporal regulation of developmentally important genes by noncoding RNAs, laying the foundation for a vast research field exploring the critical role of miRNAs in healthy and malignant biology. miRNAs regulate gene expression post-transcriptionally through the complementary binding of 5 to 6 base pair seed sequences to target messenger RNA (mRNA), leading to transcript degradation or the inhibition of translation,³ with ongoing work investigating noncanonical roles.⁴ The nuanced and highly conserved nature⁵ of miRNA-mediated gene regulation has long been of particular interest in hematopoiesis, a temporal process by which blood cells are formed from hematopoietic stem cells (HSCs). The stepwise commitment of stem and progenitor cells to mature cell lineages is controlled by networks of transcription factors, signaling cascades, and epigenetic modifications that are tightly regulated to orchestrate differentiation while maintaining the requisite pools of precursor cells.⁶ Consistent findings have demonstrated the critical role of miRNAs for maintaining and controlling hematopoiesis,^7-10 offering valuable insights into both normal blood cell development and hematological disorders. However, progress in this research area has been hindered by concerns with replicability,¹¹ and with the functional translation¹² of published results. Here, we discuss common challenges in miRNA research in malignant hematology, including the commonly overlooked complexity of miRNA-mRNA interactions, the lack of standardized methodologies, and frequent over-simplification of reported findings. We highlight key miRNA-driven discoveries in leukemias, and propose strategies to address concerns about their reliability, advocating for their continued investigation.

The spatial and temporal control of HSC development is critical for healthy cell maturation,^13-15 and the aberrant regulation of miRNA-mRNA interactions, such as with miR-146a^16-18 and miR-126,¹⁹ has been extensively linked to hematological malignancies, including acute myeloid leukemia (AML).^20-26 While the roles of integral transcription factors in hematology, such as TAL1,²⁷ RUNX1,²⁸ GATA1,²⁹ and MYB,³⁰ have been parsed out through miRNAs, the complexity of miRNA-mRNA interactions is commonly understated, and the restoration of dysregulated pathways continues to be an elusive therapeutic goal.³¹ During processing into the RNA-induced silencing complex (RISC), miRNA-hairpins are cleaved and strand selection retains the less thermodynamically stable strand to incorporate as the guide strand, while the passenger strand is frequently, but not always, degraded.^32-35 The seed sequences of both strands confer unique pleiotropic effects through the targeting of hundreds or thousands of mRNAs, thus the stability of RISC effectively controls an extensive network of downstream interactions, which are challenging to map comprehensively.^32,33,36-38 Studies drawing conclusions about a dominant strand while neglecting possible regulatory effects of the other strand, or of sequence variations (isomiRs),³⁸ may lead to the misinterpretation of results and incomplete conclusions about miRNA function. For example, comprehensive analyses revealed that both strands of the miR-223 duplex actively regulate myeloid differentiation and leukemia progression, highlighting the functional complexity of miRNA-mediated regulation.³³ Similarly, dysregulated gene networks can directly perturb expression patterns of miRNAs, as seen in the TP53 transactivation of the miR-15/miR-16 cluster in chronic lymphoid leukemia.³⁹ 

In the same way, the binding potential of the miRNA-guided RISC to a target mRNA will be heavily influenced by local biochemical conditions,⁴⁰ and by the abundance of a given miRNA or mRNA transcript.⁴¹ While in silico prediction software such as TargetScan⁴² and miRDB⁴³ have provided great insights to identifying interactions by assessing the complementary sequences and thermodynamic binding potential of miRNAs to mRNA targets, the direct binding is further influenced by other cell intrinsic factors. For example, miR-155 is upregulated in AML, but its effect on leukemic maintenance differs by disease subtype, suggesting added layers of regulation in this axis that remain unresolved between disease subtypes.^44,45 This additional regulation can include post-transcriptional modifications, such as N⁶-methyladenosine, which is mediated by RNA-binding proteins (RBPs) like FTO, and can affect the stability or structure of long noncoding RNAs and mRNAs, as well as miRNA biogenesis.^46,47 This modification is often mediated through the specific subcellular localization of interactors and cofactors, and they have been extensively linked to the maintenance of hematopoiesis and to AML.^46-53 Further, RBPs such as Lin28⁵⁴ and DICER-TRBP⁵⁵ can modulate miRNA-RISC interactions by providing steric hindrance, stabilizing target mRNAs, or altering the conformation of the mRNA 3′ UTR, thus affecting miRNA accessibility.⁵⁶ Notably, the polyadenylation complex CCR4-NOT, which mediates the shortening of mRNAs in the initiation of decay, is recruited to target mRNAs through interaction with RISC and RBPs, with evidence showing the noncatalytic subunit CNOT3 to be essential for AML.^57,58 Although these regulatory elements significantly influence miRNA-mRNA interactions, they are frequently neglected due to the regulatory complexity in experimental designs. Conventional assays, such as luciferase reporters employing truncated segments of 3′ UTRs tested in heterologous cell lines, fail to replicate these conditions adequately. Thus, more physiologically relevant models and sophisticated experimental approaches are necessary to fully elucidate biologically meaningful interactions. The complexity of miRNA targeting is further amplified by context-specific regulation, which is dependent on cellular state, developmental stage, and microenvironmental stress.³ Tissue-specific dynamics introduce another layer of complexity,⁵⁹ as interactions identified in one cell type may not be conserved in others. For example, diverse context-specific functions have been defined for miR-125b, including an oncogenic role in numerous cancers such as glioblastoma,⁶⁰ and a tumor-suppressive role in cancers like esophageal squamous cell carcinoma.⁶¹ In hematopoiesis, the numerous cell types and tightly regulated transcript expression profiles create significant research hurdles. This is exemplified by the regulation of GATA2 during HSC self-renewal, progenitor differentiation, and terminal megakaryocytic differentiation.^62,63 Such regulatory precision extends to malignant cells, where miRNA-mRNA interactions differ between leukemia subtypes and even disease stages. For instance, miR-708 exhibits contrasting regulatory roles within the same leukemia model depending on cellular context and cofactors.⁶⁴ 

Since the initial characterization of depleted miR-15a and miR-16-1 expression in chronic lymphoid leukemia,⁶⁵ numerous miRNAs have been implicated in dysregulated hematopoiesis and leukemias. However, the translation of these findings into effective therapeutic strategies remains limited in hematology.^12,66 Promising results have demonstrated the therapeutic potential of miRNA-based treatments in other cancers in vivo, such as the antitumor effects of lipid nanoparticle encapsulated miR-193a-3p (INT-1B3) in metastatic breast and liver cancer cells.⁶⁷ Further, in hematological malignancies, exciting examples like oligonucleotide inhibitor of miR-155 have shown promise in activated B-cell subtype of diffuse large B-cell lymphoma,⁶⁸ but clinical translation as a whole remains challenging. A major barrier is the over reliance on reductionist models, such as immortalized cell lines, that fail to replicate the complexity of the human bone marrow microenvironment. These models lack the spatial organization, heterogeneity, and niche-specific cues necessary to fully capture the dynamic regulation of miRNA-mRNA interactions.⁶⁹ For example, even within ostensibly identical AML cell lines harboring the KMT2A-MLLT3 translocation (t(9;11)), transcriptomic analyses have revealed vast interline variability in transcript expression profiles, challenging the assumption of functional equivalence across models.⁷⁰ In vivo models have better reflected the complexity of the bone marrow microenvironment and developmental biology, and have enabled foundational discoveries in hematology such as the maintenance of the HSC pool via miR-127 in a murine model,⁷¹ or the cell fate decision between erythroid and megakaryocytic lineages mediated by miR-126 and miR-150 shown in zebrafish.⁷² Additionally, in vivo xenograft models have led to more nuanced findings, such as expanding on the described role of miR-223 from Notch signaling⁷³ to elucidate its function in promoting lineage commitment, differentiation, and the modulation of AML.⁷⁴ In this context, hematopoietic organoids represent a significant methodological advancement. As 3-dimensional, stem cell–derived systems that emulate the structural and cellular complexity of the bone marrow niche, they allow for controlled, lineage-specific tracking of miRNA activity during hematopoietic differentiation and leukemogenesis.⁷⁵ While miRNA applications in organoid systems are already being explored in diseases like colorectal cancer,⁷⁶ their adoption in hematology remains nascent. Excitingly, recent studies have begun using vascularized bone marrow organoids to recapitulate hematopoietic hierarchies and cytokine gradients.^75,77 Incorporating such models offers a powerful opportunity to dissect context-specific miRNA functions in a physiologically relevant model emulating microenvironment and immune cell interactions beyond what is possible in immunocompromised murine models. This is critical to fully characterize regulators like miR-126, which maintains leukemic stem cells and chemotherapy resistance in chronic myeloid leukemia through interactions with endothelial cells,⁷⁸ and miR-125b, which maintains the multiple myeloma microenvironment through interleukin-6R.⁷⁹ Moving forward, leveraging hematopoietic organoids alongside traditional models will enable researchers to ask more targeted and translationally meaningful questions, improving both the reproducibility of findings and their therapeutic relevance. Rather than oversimplifying biology to fit our tools, we can now adapt our tools to meet the complexity of the system.

Although a lack of standardized methodologies still poses challenges for reproducibility and clinical translation in miRNA research in hematology, researchers are actively working to improve and unify experimental approaches. Over the past 3 decades, advancements in biotechnology and refined functional assays have significantly enhanced the precision and depth of scientific inquiry. It is important to recognize that the limitations of earlier studies often stemmed from the practical and financial constraints associated with emerging technologies at the time. This evolution represents progress in this field of research, although it inevitably introduces variation between research groups that reflect contemporary standards. Furthermore, the inherent complexity of the hematopoietic hierarchy and the diverse cellular contexts significantly contribute to the variability in miRNA expression profiles reported across studies. This variability reflects the broader complexity intrinsic to hematology research, which is often mistakenly attributed solely to miRNAs. For example, comparing peripheral blood-derived HSCs to those isolated from bone marrow aspirates can introduce significant bias due to immunological and functional differences resulting from distinct microenvironmental influences, which is compounded by differences in miRNA stability during sample processing.⁸⁰ Seemingly minor procedural variations, such as the timing of blood sampling, centrifugation steps, or storage conditions, can profoundly affect miRNA integrity, skewing expression data and obscuring biologically meaningful differences.^81-84 Due to their small size and lack of poly(A) tail, mature miRNAs are often extracted and purified separately from mRNA, amplifying the risk of procedural variations.⁸³ Further, varying analysis thresholds and normalization strategies exacerbate these differences between research groups. Increased uniformity has been achieved through the minimum information for publication of quantitative real-time polymerase chain reaction (PCR) experiments⁸⁵ and the International Organization for Standardization,⁸⁶ and will serve to overcome these methodological inconsistencies that may otherwise create fragmented interpretations that hinder our understanding of miRNA roles in disease initiation and progression.

Although biobanking initiatives represent significant progress by enabling retrospective analysis and functional assays from stored patient samples, variability in privacy protocols, incomplete clinical annotations, and inconsistent handling continue to impede broad epidemiological studies and robust population-level investigations.⁸⁶ Increased emphasis on data annotation and processing has enabled leukemic patient miRNA expression repositories like the Cancer Genome Atlas Program (TCGA) to harmonize to the latest human reference genome while maintaining high concordance with legacy data. The integrity of these data has led to the introduction of prognostic miRNA biomarker scores in AML,⁸⁷ while providing a workflow for groups to adopt for their own analysis and data deposit to the Gene Expression Omnibus. The utility in combining high-quality data can be exemplified by the novel identification of miR-181a/b as a prognostic marker in hematological malignancies.⁸⁸ However, commonly used bulk-sequencing approaches have notable limitations, including inadequate sensitivity for lowly expressed miRNAs or minor cell subsets, confounding accurate characterization of miRNA-driven regulatory networks.⁸⁹ Recent advancements such as single-cell sequencing and spatial transcriptomics have improved resolution, and enabled the identification of clinically relevant miRNA-mRNA interactions like miR-125a repressing histone deacetylase 6 to promote chemoresistance in a subset of patients with KMT2A-rearranged AML,⁹⁰ though these methods remain constrained by cost, replicability, and limited sequencing depth.

To address these challenges, the field must prioritize standardization and methodological transparency. Specifically, we advocate for rigorous documentation and public sharing of sample harvesting, storage, and experimental protocols. Further, integrating sensitive validation methods such as digital droplet PCR or quantitative real-time PCR with functional assays will improve replicability and translational potential of findings. Standard practice to confirm functional relevance should evolve to include sophisticated validation techniques such as crosslinking immunoprecipitation⁹¹ or CRISPR-based miRNA-activated genome editing.⁹² An increased emphasis on standardized and validated methodologies, complemented by improved collaboration between academia, clinical pathology, and industry, will significantly enhance the reproducibility, reliability, and clinical relevance of miRNA research outcomes in hematology.

A barrier in miRNA research within hematology is the frequent oversimplification of inherently complex and context-specific interactions. Studies often promote straightforward narratives, prioritizing neatly packaged conclusions that align conveniently with existing hypotheses or expectations. However, this approach underestimates the biological nuance of miRNA regulation, frequently dismissing contradictory or ambiguous findings as methodological errors or irrelevant data.¹² For instance, investigations that highlight a single miRNA-target relationship without considering broader network interactions or differences in profiling approaches can obscure biologically meaningful insights and reduce confidence in conclusions when conflicting results inevitably emerge, as with the initially ambiguous role of miR-125b in cancer.⁹³ 

The practice of simplification hinders scientific progress by preventing full appreciation of miRNA-mediated regulatory networks in the context of heterogeneous cell populations and disease states. As previously noted, the inherent complexity of regulatory networks in the hematopoietic system introduces inherent variability to this research field, which is frequently misattributed to miRNAs. The limitations of current models, such as immortalized cell lines that lack niche interactions with characterized roles in disease,⁶⁹ or patient-derived samples that often represent mixed cell states,^80,94 necessitate embracing complexity and transparently reporting conflicting results to drive meaningful progress. For example, the characterization of miRNA expression profiles as a diagnostic for specific molecular phenotypes in t(14;18)-negative follicular lymphoma represents a direct translational benefit achieved in a patient subset that otherwise lacked known molecular features.⁹⁵ Future clinical relevance depends on recognizing and openly discussing the nuances and apparent contradictions that arise in miRNA studies, rather than selectively presenting results that merely confirm initial hypotheses.

Excitingly, studies that have embraced complexity by defining specific contexts or mechanisms continue to significantly advance our understanding. For example, recent findings identifying decreased MEG3 and miR-493-5p in cytarabine-resistant AML cells have revealed a novel therapeutic vulnerability via METTL3-mediated N⁶-methyladenosine modulation of MYC expression.^96,97 Similarly, work defining the mitochondrial activation of the proto-oncogenic miR-106a-363 cluster specifically in adverse-risk AML demonstrates how clarifying cellular context can reveal distinct therapeutic opportunities.⁹⁸ These studies highlight the benefits of careful, phenotype-driven miRNA investigations in specific cellular contexts, which can provide strong, functionally validated frameworks for future research.

The precise regulation of hematopoiesis by miRNAs highlights their significant scientific and therapeutic potential. Yet, biological complexity, perceived methodological inconsistencies, and a tendency toward oversimplification continue to impede progress. Overcoming these challenges will require transparent reporting, standardized workflows, and rigorous functional validation. Emerging tools, such as single-cell sequencing, spatial transcriptomics, hematopoietic organoids, and CRISPR-based assays, offer promising avenues to dissect miRNA biology in physiologically relevant contexts. Optimizing these pipelines and harnessing targeted delivery systems,^31,99 including lipid nanoparticles of miRs or anti-miRs,¹⁰⁰ will further accelerate clinical translation. Notably, recent phenotype-driven studies exemplify the value of embracing complexity, setting the stage for future discoveries. With continued methodological refinement, miRNA research in hematology is well positioned to uncover therapeutic vulnerabilities and drive transformative clinical advances.

F.K. was supported by grants from the Leukemia Lymphoma Society of Canada and Michael Smith Health Research BC.

Contribution: All authors discussed the concepts and scope of the manuscript, analyzed relevant literature, wrote the manuscript, and reviewed and approved the final version of the manuscript.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Florian Kuchenbauer, Terry Fox Laboratory, BC Cancer Research Institute, 675 West 10th Ave, Vancouver, BC V5Z 1L3, Canada; email: fkuchenbauer@bccrc.ca.